These are chat archives for rdfjs/public

5th
Oct 2016
Tim Berners-Lee
@timbl
Oct 05 2016 13:35
Sandro? re "Not sure what you mean about getting into a twist.” — not worth getting your knickers in a twost over (the differnce between a graph value and a graph variable) — when one gets upset or crazy over something most people think is unimportant
If you get your panties in a twist it gets hard to make progress
</analogyElaboration>
Having to use the terms Gbox and GSnap instead of graph variable and graph seems to me to impede progress
Sandro Hawke
@sandhawke
Oct 05 2016 15:45
Like a lot of these things, it doesn't matter until it does. I'm not the one bringing this up, at all. But saying "just use my terminology" (as you just did) doesn't work, because... empirically... people won't. Specifically, everybody says "graph" when they mean "graph variable" (gbox), and telling them not to doesn't help.
Dmitri Zagidulin
@dmitrizagidulin
Oct 05 2016 15:50
the rdfjs taskforce peoples are trying to standardize on 'dataset' for gsnap (if I understand the distinctions correctly)
fwiw.
Sandro Hawke
@sandhawke
Oct 05 2016 15:55
That would contradict both sparql and rdf, so I trust you're misunderstanding.
Dmitri Zagidulin
@dmitrizagidulin
Oct 05 2016 15:55
ok, yeah, most likely :)
tho in the interest of further education - what do sparql & rdf say?
Tim Berners-Lee
@timbl
Oct 05 2016 17:10
I’m not saying use my terminology as much as “use yoru existing"
If you have a language with var and const then you can say
var g1
const g2 = { :fred a :Person }
var g1
const  fact = {  :fred a :Person }
var g3 = fact
g3   +=  {  :fred a :Male }
global commonKnowledge
g3 += commonKnowledge
say, where the computer language has the ideas of variable and valeus built in. We don’t have pins and angels discussion over strings and arrays — why do we have to for graphs?
Tim Berners-Lee
@timbl
Oct 05 2016 17:16
Maybe it is because people are used to oprogramming languages but they are not used to KR langauages, which are different
Dmitri Zagidulin
@dmitrizagidulin
Oct 05 2016 17:17
what are KR languages?
elf Pavlik
@elf-pavlik
Oct 05 2016 17:19
does using IRI as ‘the variable’ which references a graph ‘the value’ assumes that graph name denotes the graph? @sandhawke already mentioned on this channel that he considers this note in RDF 1.1 spec a process glitch https://www.w3.org/TR/rdf11-concepts/#h_note_4
Despite the use of the word “name” in “named graph”, the graph name is not required to denote the graph. It is merely syntactically paired with the graph. RDF does not place any formal restrictions on what resource the graph name may denote, nor on the relationship between that resource and the graph. A discussion of different RDF dataset semantics can be found in [RDF11-DATASETS].
Tim Berners-Lee
@timbl
Oct 05 2016 17:21
Knolwedge representation languages, like RDf, cyc, etc etc https://en.wikipedia.org/wiki/Knowledge_representation_and_reasoning
Dmitri Zagidulin
@dmitrizagidulin
Oct 05 2016 17:22
ahhh i see
Tim Berners-Lee
@timbl
Oct 05 2016 17:23
“”"After CycL, a number of ontology languages have been developed. Most are declarative languages, and are either frame languages, or are based on first-order logic. Modularity—the ability to define boundaries around specific domains and problem spaces—is essential for these languages because as stated by Tom Gruber, "Every ontology is a treaty- a social agreement among people with common motive in sharing.” “""
With linked data we de-emphasize the logic and inference, and stick to the basic data layer
Sandro Hawke
@sandhawke
Oct 05 2016 18:37
@timbl I don't think many languages have the right terminology. For instance, immutability rarely is transitive via pointers. A 'const' in JS only goes one step. Also, immutability and identity are usually separate aspects, in programming languages. In my mind, this should be a concept people can use when they need to disambiguate, but can normally be ignored.
@dmitrizagidulin RDF and SPARQL says "graph" and "dataset" are pure mathematical (immutable, identical) concepts, aka g-snaps (or sets of g-snaps, for a dataset). In practice, people rarely mean them that way. Even the SPARQL language itself doesn't use the term with this precision.

Example 2 from https://www.w3.org/TR/sparql11-update/#insertData INSERT DATA { GRAPH <http://example/bookStore> { <http://example/book1> ns:price 42 } }

That use of GRAPH is clearly to refer to a mutable structure (a g-box)

Sandro Hawke
@sandhawke
Oct 05 2016 20:54
By "identical" I mean "identical if equal". If they have the same triples in them, they must have identical metadata.
(You can't say you fetched one at 2pm and the other at 3pm, since they are the same thing.)